The Year AI Got Booed at Graduation
Somewhere this spring, a commencement speaker closed out a speech with some inspirational remarks about artificial intelligence. The future. The opportunity. The moment we are living through. The crowd's response: sustained booing.
Not polite skepticism. Not quiet discomfort. Actual booing. From the graduating class supposedly about to inherit all this magnificent opportunity.
That story has been stuck in my head for weeks. Because the people doing the booing were 22 years old. They were the demographic most fluent with this technology, the ones who've been using ChatGPT to write their essays for three years, the cohort that tech executives are constantly describing as the future workforce. And they heard a pitch for AI at their graduation ceremony and they booed it.
That should mean something.
Meanwhile, in San Francisco: celebration. Google I/O arrived in May like a product dump in a trenchcoat. Over a dozen announcements. New Gemini models. A 24/7 background agent called Gemini Spark that lives on your phone and autonomously handles event planning and calendar management. Shopping integrations across every Google surface. YouTube AI search. A price cut on the $250/month subscription down to $100. The overall vibe: we have won. Everything is going extremely well. Watch these demos.
The week before: Anthropic released new API features. OpenAI won its lawsuit against Elon Musk. The industry's collective mood has been somewhere between triumph and mild impatience that the world hasn't caught up yet.
They might want to check the polls.
Here is what a few surveys found this month:
More than 70 percent of Americans believe AI is advancing too quickly, according to an Economist/YouGov survey. This is bipartisan: 68% of Republicans and 77% of Democrats agree. The two groups disagree on almost everything else right now. They found consensus here.
Only 18 percent of young people ages 14 to 29 say they feel hopeful about AI, according to Gallup. Not worried, not neutral. Hopeful. Eighteen percent. The other 82 percent aren't carrying the utopian vision the industry keeps projecting onto them.
A majority of registered voters, 57 percent, say AI risks outweigh its benefits. Twenty-six percent hold positive views of AI overall. Forty-six percent hold negative ones.
The Harris Poll 100, which measures corporate reputation, found several prominent AI companies now rank below US Immigration and Customs Enforcement in public perception. That is not a misprint. ICE. The agency anchored at the bottom of reputation surveys for years running. AI companies looked at that bar and went under it.
When one AI CEO was presented with this data, the response was: "We don't really see that."
That quote deserves to be laminated and displayed in the lobby of every AI company in the Bay Area.
What's driving this? It's not technophobia. The public isn't afraid of the future in the abstract. People are responding to a specific list of concrete concerns, and none of them are irrational.
Jobs. Nearly three in five workers told Gallup they're worried about AI's impact on their careers. Not gig workers: white-collar professionals. Writers, analysts, customer service roles, people who built careers specifically designed to be automation-resistant. They are watching those careers get repriced in real time.
Energy. AI data centers are projected to consume 15 to 25 percent of US electricity by 2030. The grid was not built for this load. In the first quarter of 2026, a record number of data center projects were canceled due to community opposition. Morgan Stanley analysts described it as a "binding constraint" on AI expansion. Communities said no by preventing construction from starting.
Wealth. The benefits of AI are concentrating in a narrow group of companies and investors. The costs, the energy draw, the displaced workers, the training data scraped without clear consent, are diffuse and widely distributed. People have noticed the asymmetry.
None of this is a misunderstanding that better messaging will fix.
The industry's standard response is: wait. Wait until the products get better. Wait until the benefits materialize. Wait until people see what AI can do for them personally. The implicit premise is that public opinion will eventually converge with technological reality once the value becomes undeniable.
Maybe. But that argument requires a kind of patience from the public that the industry isn't extending in return. The companies that built these systems decided how fast to move, how benefits would flow, whether to be transparent about tradeoffs. They did not consult the graduating class. They moved fast. They are now surprised the graduating class has opinions about the results.
You can't engineer your way out of a consent problem. You can't ship a product that resolves the underlying concern, which isn't "the model isn't good enough" but "I didn't agree to this and I don't trust who's running it."
None of this means AI isn't significant. It clearly is. The productivity gains in certain domains are real. Some of those Google I/O demos were genuinely impressive. If you want to understand what AI tools actually do and how real people are using them, chatbot.gallery has profiles of hundreds of tools across every category, a useful reality check on what's actually shipping versus what's being announced.
But there's a version of the next five years where the industry's defining challenge isn't model capabilities. It's whether enough people outside the industry trust what's being built to allow it to continue expanding. Data center cancellations suggest that's not a hypothetical.
The commencement speaker probably gave a good speech. They usually do. But the graduates booing weren't rejecting progress. They were pushing back on a specific pitch from an industry that has spent years describing them as beneficiaries without asking them much.
That seems like a reasonable distinction. The industry might consider engaging with it before next graduation season.
